Anand Socket-F vs. Woodcrest

[sandmanwn]

http://www.anandtech.com/IT/showdoc.aspx?i=2892

http://www.anandtech.com/IT/showdoc.aspx?i=2892

Woodcrest performs better near saturation point. Although who would run a server at saturation point anyway.

Opteron does better on power consumption. Surprisingly.

Good to finally see a test between the two that wasnt based on encoding/decoding dvd/mp3s or whatever other synthetic crap that has little relevance to real world applications.

 

[apt403]

yeah, i just read it about a hour ago, the wierd part is that i was more interested in the case that the systems where housed in then the systems or benchmarks.

 

[Pippero]

Hooray for Anandtech, which benchmarks server CPUs with server tests! :trophy:
Seriously, Tom's HW is soooo falling behind these days (years?)...

 

[sandmanwn]

the wind tunnel design on the xeons is pretty common. i wonder if its a requested design from intel?

ive purchased a lot HP ML's and DL's from G1's to G4's for both AMD and Xeon and the wind tunnel has been the design of every intel solution.

 

[1Tanker]

http://www.anandtech.com/IT/showdoc.aspx?i=2892

http://www.anandtech.com/IT/showdoc.aspx?i=2892

Woodcrest performs better near saturation point. Although who would run a server at saturation point anyway.

Opteron does better on power consumption. Surprisingly.

Good to finally see a test between the two that wasnt based on encoding/decoding dvd/mp3s or whatever other synthetic crap that has little relevance to real world applications.I hope Intel doesn't regret going with FB-DIMM's. :?

 

[sandmanwn]

indeed, it seems to be squandering their power/performance ratio

still trying to figure out why intel is so insistent on shaking things up in the memory market. this is really one area they need to just backup and go with the flow.

 

[weskurtz81]

Maybe they have stakes in the FB-Dimm tech/market..... just a thought.

wes

 

[Barcelona_Xtreme]

Good review from Anand.

Rev-F Opterons do have true potential against intel's Woodcrest and Cloverton. I guess barcelona will widen the gap even more putting it as the clear leader in 2P-4P-8P servers and HPCs.

When would we see such a review from THG? :?

 

[sandmanwn]

good question. as mentioned in the article Anand actually sent a request to AMD for a preview.

AMD is having a difficult time with meeting supply/demand. I would guess THG would have to make a similar request to get a preview.

 

[Ycon]

They didnt test 2220 vs. X5355 (~same power consumption for the chip itself)
While having done some not-so-bad tests in the last months, AnandTech is turning to AMDTech again...

 

[weskurtz81]

ycon,

I don't really understand what you are getting at with the first part of your statement, please elaborate...

wes

Edit: Oh I get it, you were being a "fanpoi". I think that is how you decided to spell it. Anyway, grow up, get a life.... know one cares man. This is a good review, just deal with it, stop being another 9inch.

 

[Barcelona_Xtreme]

Just wanted to contribute to your thread with this other benchmark:

Cloverton can't keep up with Opteron

 

[dasickninja]

You do realize that the only time the 2P Xeons got beat by the Opterons were when it was a 4P Opteron. Try reading the entire test the next time.

 

[Barcelona_Xtreme]

You do realize that the only time the 2P Xeons got beat by the Opterons were when it was a 4P Opteron. Try reading the entire test the next time.

LOL. :lol:
We're talking about 4 processors vs 4 processors, don't matter if it's dual-quad core or quad-dual cores. :lol:
Even if intel had Cloverton in a 4-way setup they would be doing injury to such a system thanks to the FSB bottlenecking.

Even a 2P Opteron performs better than Woodcrest in SPECfp and battles its way in the other benchmarks.

Anyhow, judging by your reasoning, Quad-FX is not in intel's kentsfield lands?

 

[accord99]

Just wanted to contribute to your thread with this other benchmark:

Cloverton can't keep up with Opteron

From those tests, it's also the case that Opteron can't keep up with Tulsa.

 

[Action_Man]

Just wanted to contribute to your thread with this other benchmark:

Cloverton can't keep up with Opteron

9nm! You're back!

 

[heartview]

Just wanted to contribute to your thread with this other benchmark:

Cloverton can't keep up with Opteron

Only one small error in that one about the 4004... it didn't have 640KB of RAM, it was only 640 BYTES. I don't remember which of their processors went above 64k first but it was somewhere around the 8080.

EDIT: some corrections.

EDIT: Ok, after some homework, the 8086 was the first to break the 64k barrier (at least from Intel).

 

[samxxxii]

http://www.anandtech.com/IT/showdoc.aspx?i=2892

http://www.anandtech.com/IT/showdoc.aspx?i=2892

Woodcrest performs better near saturation point. Although who would run a server at saturation point anyway.

Opteron does better on power consumption. Surprisingly.

Good to finally see a test between the two that wasnt based on encoding/decoding dvd/mp3s or whatever other synthetic crap that has little relevance to real world applications.

Actually there are big companies that are using VMware....! in orther to extract all the juice from their servers....!!! instead of geting 20%-30% of load you can get around 80% + Load....! and the benefit from a more stable plataform in cases where you've got 1 server down you can actually manage them with your other servers and at the same time you are fixing the one who is down...!!!

Hurra Virtual Machines....!!!

Cheers

Sam